This document was generated from CDN thread
Created by: Mert Guray on 03-06-2011 10:43:18 AM
Hello everyone,
Does anybody use any kind of data like expected wait time of the customer waiting in a queue.
We are trying to prompt the massage like "You've got 3 mins to wait. Please hold the line..."
Does cisco supply such a metric. Or thhere maybe someone who makes a calculation using other metrics of ICM Script.
We are using ICM 7.5.
Thanks .
Subject: RE: New Message from Mert Guray in Customer Voice Portal (CVP) - General Di
Replied by: GEOFFREY THOMPSON on 03-06-2011 10:58:09 AM
Of course. Look up Expected Wait Time (EWT) calculations in the ICM literature.
Regards,
Geoff
Subject: Re: New Message from Mert Guray in Customer Voice Portal (CVP) - General Di
Replied by: Janine Graves on 03-06-2011 11:01:09 AM
This is going to depend upon your queuing, etc. The EWT that is used for
Courtesy Callback ICM script is this:
((SkillGroup.%1%.RouterCallsQNow+1) -
(ValidValue(SkillGroup.%1%.AvgHandledCallsTimeTo5, 20))
/ Max ( SkillGroup.%1%.Ready,
(SkillGroup.%1%.TalkingIn + SkillGroup.%1%.TalkingOut
+SkillGroup.%1%.TalkingOther))))
On 6/3/2011 10:43 AM, Cisco Developer Community Forums wrote:
> Mert Guray has created a new message in the forum "General Discussion
> - All Versions":
>
> --------------------------------------------------------------
> Hello everyone,
> Does anybody use any kind of data like expected wait time of the
> customer waiting in a queue.
> We are trying to prompt the massage like "You've got 3 mins to wait.
> Please hold the line..."
> Does cisco supply such a metric. Or thhere maybe someone who makes a
> calculation using other metrics of ICM Script.
> We are using ICM 7.5.
> Thanks .
> --
> To respond to this post, please click the following link:
>
> <http://developer.cisco.com/web/cvp/forums/-/message_boards/view_message/3999039>
>
> or simply reply to this email.
--
Janine Graves
Subject: RE: Expected Wait Time in ICM Script
Replied by: Mert Guray on 08-06-2011 08:02:24 AM
Thanks all,
It seems that there is no easy value to use, according to the documentation. The first sentence goes like this:
The Script Editor's built-in Minimum Expected Delay (MED) calculation does not apply to IPCC. Instead, you can use a formula ......
Janine, thanks for your example formula, but i think there must me a multiply sign instead of minus. Do you agree?
<font face="Times New Roman" size="2"><font face="Times New Roman" size="2">
((SkillGroup.%1%.RouterCallsQNow+1) -
(ValidValue(SkillGroup.%1%.AvgHandledCallsTimeTo5, 20))
/ Max ( SkillGroup.%1%.Ready, (SkillGroup.%1%.TalkingIn + SkillGroup.%1%.TalkingOut +SkillGroup.%1%.TalkingOther))))
</font></font><font face="Times New Roman" size="2">
</font>
Subject: RE: Expected Wait Time in ICM Script
Replied by: Bill Webb on 08-06-2011 08:56:50 AM
No, the multiply sign is correct. Basically what the formula is saying is:
Take the number of calls currently waiting in queue and multiply them with the average amount of time it takes to handle a single call. This gives you a total time required to handle all of the calls currently in queue. You then divide that amount by the number of agents capable of handling those calls, which essentially gives you the expected amount of time that a new call entering the queue would be waiting to be answered.
- Bill
Subject: Re: New Message from Mert Guray in Customer Voice Portal (CVP) - General Di
Replied by: Janine Graves on 08-06-2011 10:22:54 AM
Hi Mert,
Must've been my typo, yes if Bill agrees that multiplication is correct,
then I agree too.
Thanks, Janine
On 6/8/2011 8:02 AM, Cisco Developer Community Forums wrote:
> Mert Guray has created a new message in the forum "General Discussion
> - All Versions":
>
> --------------------------------------------------------------
> Thanks all,
> It seems that there is no easy value to use, according to the
> documentation. The first sentence goes like this:
>
> The Script Editor's built-in Minimum Expected Delay (MED) calculation
> does not apply to IPCC. Instead, you can use a formula ......
> Janine, thanks for your example formula, but i think there must me a
> *multiply sign instead of minus*. Do you agree?
> Â
> <font face="Times New Roman" size="2"><font face="Times New Roman"
> size="2">
> ((SkillGroup.%1%.RouterCallsQNow+1) *-
> *(ValidValue(SkillGroup.%1%.AvgHandledCallsTimeTo5, 20))
> / Max ( SkillGroup.%1%.Ready, (SkillGroup.%1%.TalkingIn +
> SkillGroup.%1%.TalkingOut +SkillGroup.%1%.TalkingOther))))
> </font></font><font face="Times New Roman" size="2">
> Â
> </font>
> Â
> Â
> --
> To respond to this post, please click the following link:
>
> <http://developer.cisco.com/web/cvp/forums/-/message_boards/view_message/4021597>
>
> or simply reply to this email.
--
Janine Graves
Subject: RE: Expected Wait Time in ICM Script
Replied by: Bill Webb on 08-06-2011 10:37:53 AM
Haha - sorry, I had transposed it also, and thought the question was minus instead of multiply - sorry about that confusion!!
Time for more coffee... ;-)
- Bill
Subject: RE: Expected Wait Time in ICM Script
Replied by: Mert Guray on 09-06-2011 06:30:04 AM
No problem
I've got it. Thanks.
Subject: RE: Expected Wait Time in ICM Script
Replied by: deniz caglar on 09-08-2011 04:53:29 AM
Hi everyone,
We try this algoritm in one of our skillgroups but it seems the calculated results are not correct.
I wrote this calculated value to a CallVariable and got the seconds i wait in the queue, results are very different.
For example it calculated 52 sc but i my call was handled in 19 sc, and in the other one, it calculated 249 sc, but call was handled in 67sc.
I have two points now.
I think this algoritm doesnt consider the order of the queue. I mean if i am in the queue for 10 minutes and another customer is in the queue for 1 one minute, this algoritm reads the same value to both of us. We dont only read this value to the customer at the first time he is in the queue, but we read it periodically may be once a minute. So may be in the second interval of the second customer that means customer is in the queue for 1 minute, and the tenth interval of the first customer that means he is waiting for 10 minutes, both of them will be informed with the same wait time value. But the first customer will be handled before the second one.
And the other point is; in the Networkers 2011 last month, in the CVP session, they said that there would be a fix in 8.5 for Courtesy Callback. If i am not wrong, in 8.0 if the customer doesnt accept the callback option, then he will be in the last order of the queue, so they dont calculate the order of the customer in the queue dynamically every time. This issue also makes me suspicious about this algoritm.
Do you have any idea or do you use this algoritm in production for giving information to your customers?
Thanks,
Deniz
Subject: RE: Expected Wait Time in ICM Script
Replied by: Paul Antinori on 30-08-2011 04:47:48 PM
The problem we saw in the 8.0 release with EWT is that the callback time is fixed at the time when the caller initially enters the queue.
That implementation does not dynamically update the callback time when the call center call handle rate changes, which renders the ccb feature less effective especially when the callback time is long.
The enhancement in 8.5 release which is available as a patch, is that we can update the EWT (dequeue rate) for ALL of the calls in the CCB queue instead of for just the one call.
The goal of the enhancement was to get the "average callback queue time" after a callback within an agreed service level (default is 1 min).
We do not want to call back callers too early so that they would need to wait in queue for a long time.
We also do not want to call back too late so that agents may be waiting.
So we try get a large percentage of the calls handled within the service level SLA, and the remainder of the calls are falling outside this service level SLA, like a bell curve.
The randomness of call handle time is outside the control of CVP, of course. It could be why you are seeing odd results with your testing.
CVP customers could reduce the randomness of call handle time through business analysis such as separating calls for sales and technical supports into different queues.
Subject: Re: New Message from Paul Antinori in Customer Voice Portal (CVP) - General
Replied by: Janine Graves on 30-08-2011 10:24:54 PM
Hi Paul,
For purposes of recalculating EWT, what is Cisco using as the formula
for EWT while a caller is waiting for callback? And is there any way to
modify this formula?
Thanks, Janine
On 8/30/2011 4:47 PM, Cisco Developer Community Forums wrote:
> Paul Antinori has created a new message in the forum "General
> Discussion - All Versions":
>
> --------------------------------------------------------------
> The problem we saw in the 8.0 release with EWT is that the callback
> time is fixed at the time when the caller initially enters the queue.
> That implementation does not dynamically update the callback time when
> the call center call handle rate changes, which renders the ccb
> feature less effective especially when the callback time is long.
> The enhancement in 8.5 release which is available as a patch, is that
> we can update the EWT (dequeue rate) for ALL of the calls in the CCB
> queue instead of for just the one call.
>
> The goal of the enhancement was to get the "average callback queue
> time" after a callback within an agreed service level (default is 1 min).
> We do not want to call back callers too early so that they would need
> to wait in queue for a long time.
> We also do not want to call back too late so that agents may be waiting.
> So we try get a large percentage of the calls handled within the
> service level SLA, and the remainder of the calls are falling outside
> this service level SLA, like a bell curve.
> The randomness of call handle time is outside the control of CVP, of
> course. It could be why you are seeing odd results with your testing.
> CVP customers could reduce the randomness of call handle time through
> business analysis such as separating calls for sales and technical
> supports into different queues.
> --
> To respond to this post, please click the following link:
>
> <http://developer.cisco.com/web/cvp/forums/-/message_boards/view_message/4431792>
>
> or simply reply to this email.
--
Janine Graves
Subject: RE: Expected Wait Time in ICM Script
Replied by: Xin Yang on 30-08-2011 03:35:13 PM
Hi Deniz,
This is interesting. How about minus current queue length to get the EWT in middle of the queue?
Thanks
James
Subject: RE: Expected Wait Time in ICM Script
Replied by: Paul Antinori on 31-08-2011 08:55:10 AM
Hi Janine,
The default ICM script for CCB has the exact same formula as above. Of course you can customize it for the customer call center.
ValidValue(((XXXX.RouterCallsQNow+1)*(ValidValue(XXXX.AvgHandledCallsTimeTo5,20))/max(XXXX.Ready,(XXXX.TalkingIn+XXXX.TalkingOut+XXXX.TalkingOther))),100)
For 8.5 ES it appears in 2 places in the script to allow for more frequent updating via ged125 messages of the CCB Queue EWT calculation.
Paul
Subject: RE: Expected Wait Time in ICM Script
Replied by: Sethuramalingam Balasubramanian on 03-01-2012 12:47:27 AM
Hi Paul,
Can we know in which version of UCCE/ICM this fix is available or what is the appropriate ES version to this fix?
Thanks!
-Sethu
Subject: RE: Expected Wait Time in ICM Script
Replied by: Brenda McAleenan on 12-09-2013 07:32:00 AM
Hi All
I've made a couple of amendments to this to allow consideration of available agents (without having to use another IF node). If agents are available we'll always set EWT to 0 as they'll be answered straight off.
if(SkillGroup.%1%.Avail>=1,0,ValidValue(((XXXX.RouterCallsQNow+1)*(ValidValue(XXXX.AvgHandledCallsTimeTo5,20))/
max(XXXX.Ready,(XXXX.TalkingIn+XXXX.TalkingOut+XXXX.TalkingOther))),100)
Additionally if you queue to multiple skills you can tweak the formula to suit this. First thing is that you'll only want to check RCQ in one skill (as it should be queuing to all)
if(SkillGroup.%1%_Lvl1.Avail+SkillGroup.PrePay_Lvl2.Avail>=1,0,ValidValue(((SkillGroup.%1%.RouterCallsQNow+1)*(ValidValue((ValidValue(SkillGroup.%1%.HandledCallsTimeTo5,0)+ValidValue(SkillGroup.%2%.HandledCallsTimeTo5,0))/(ValidValue(SkillGroup.%1%.CallsHandledTo5,0)+ValidValue(SkillGroup.%2%.CallsHandledTo5,0)),1))/max((SkillGroup.%1%.Ready+SkillGroup.%2%.Ready),(SkillGroup.%1%.TalkingIn+SkillGroup.%1%.TalkingOut+SkillGroup.%1%.TalkingOther+SkillGroup.%2%.TalkingIn+SkillGroup.%2%.TalkingOut+SkillGroup.%2%.TalkingOther))),100))
You can't use AvgHandledCallsTimeTo5 here anymore as you'd be averaging an average so I've replaced with the manual calculation using HandledCallsTimeTo5/CallsHandledTo 5 looking to both skills. You also need to use the ValidValue logic as these could return NULL for some of the objects (I've noticed CallsHandledTo5 definitely skews this) and therefore default to 20 from the original ValidValue in the formula. I've also found that putting them in a custom function is much easier as they can be reused or updated much easier that way.
Created by: Mert Guray on 03-06-2011 10:43:18 AM
Hello everyone,
Does anybody use any kind of data like expected wait time of the customer waiting in a queue.
We are trying to prompt the massage like "You've got 3 mins to wait. Please hold the line..."
Does cisco supply such a metric. Or thhere maybe someone who makes a calculation using other metrics of ICM Script.
We are using ICM 7.5.
Thanks .
Subject: RE: New Message from Mert Guray in Customer Voice Portal (CVP) - General Di
Replied by: GEOFFREY THOMPSON on 03-06-2011 10:58:09 AM
Of course. Look up Expected Wait Time (EWT) calculations in the ICM literature.
Regards,
Geoff
Subject: Re: New Message from Mert Guray in Customer Voice Portal (CVP) - General Di
Replied by: Janine Graves on 03-06-2011 11:01:09 AM
This is going to depend upon your queuing, etc. The EWT that is used for
Courtesy Callback ICM script is this:
((SkillGroup.%1%.RouterCallsQNow+1) -
(ValidValue(SkillGroup.%1%.AvgHandledCallsTimeTo5, 20))
/ Max ( SkillGroup.%1%.Ready,
(SkillGroup.%1%.TalkingIn + SkillGroup.%1%.TalkingOut
+SkillGroup.%1%.TalkingOther))))
On 6/3/2011 10:43 AM, Cisco Developer Community Forums wrote:
> Mert Guray has created a new message in the forum "General Discussion
> - All Versions":
>
> --------------------------------------------------------------
> Hello everyone,
> Does anybody use any kind of data like expected wait time of the
> customer waiting in a queue.
> We are trying to prompt the massage like "You've got 3 mins to wait.
> Please hold the line..."
> Does cisco supply such a metric. Or thhere maybe someone who makes a
> calculation using other metrics of ICM Script.
> We are using ICM 7.5.
> Thanks .
> --
> To respond to this post, please click the following link:
>
> <http://developer.cisco.com/web/cvp/forums/-/message_boards/view_message/3999039>
>
> or simply reply to this email.
--
Janine Graves
Subject: RE: Expected Wait Time in ICM Script
Replied by: Mert Guray on 08-06-2011 08:02:24 AM
Thanks all,
It seems that there is no easy value to use, according to the documentation. The first sentence goes like this:
The Script Editor's built-in Minimum Expected Delay (MED) calculation does not apply to IPCC. Instead, you can use a formula ......
Janine, thanks for your example formula, but i think there must me a multiply sign instead of minus. Do you agree?
<font face="Times New Roman" size="2"><font face="Times New Roman" size="2">
((SkillGroup.%1%.RouterCallsQNow+1) -
(ValidValue(SkillGroup.%1%.AvgHandledCallsTimeTo5, 20))
/ Max ( SkillGroup.%1%.Ready, (SkillGroup.%1%.TalkingIn + SkillGroup.%1%.TalkingOut +SkillGroup.%1%.TalkingOther))))
</font></font><font face="Times New Roman" size="2">
</font>
Subject: RE: Expected Wait Time in ICM Script
Replied by: Bill Webb on 08-06-2011 08:56:50 AM
No, the multiply sign is correct. Basically what the formula is saying is:
Take the number of calls currently waiting in queue and multiply them with the average amount of time it takes to handle a single call. This gives you a total time required to handle all of the calls currently in queue. You then divide that amount by the number of agents capable of handling those calls, which essentially gives you the expected amount of time that a new call entering the queue would be waiting to be answered.
- Bill
Subject: Re: New Message from Mert Guray in Customer Voice Portal (CVP) - General Di
Replied by: Janine Graves on 08-06-2011 10:22:54 AM
Hi Mert,
Must've been my typo, yes if Bill agrees that multiplication is correct,
then I agree too.
Thanks, Janine
On 6/8/2011 8:02 AM, Cisco Developer Community Forums wrote:
> Mert Guray has created a new message in the forum "General Discussion
> - All Versions":
>
> --------------------------------------------------------------
> Thanks all,
> It seems that there is no easy value to use, according to the
> documentation. The first sentence goes like this:
>
> The Script Editor's built-in Minimum Expected Delay (MED) calculation
> does not apply to IPCC. Instead, you can use a formula ......
> Janine, thanks for your example formula, but i think there must me a
> *multiply sign instead of minus*. Do you agree?
> Â
> <font face="Times New Roman" size="2"><font face="Times New Roman"
> size="2">
> ((SkillGroup.%1%.RouterCallsQNow+1) *-
> *(ValidValue(SkillGroup.%1%.AvgHandledCallsTimeTo5, 20))
> / Max ( SkillGroup.%1%.Ready, (SkillGroup.%1%.TalkingIn +
> SkillGroup.%1%.TalkingOut +SkillGroup.%1%.TalkingOther))))
> </font></font><font face="Times New Roman" size="2">
> Â
> </font>
> Â
> Â
> --
> To respond to this post, please click the following link:
>
> <http://developer.cisco.com/web/cvp/forums/-/message_boards/view_message/4021597>
>
> or simply reply to this email.
--
Janine Graves
Subject: RE: Expected Wait Time in ICM Script
Replied by: Bill Webb on 08-06-2011 10:37:53 AM
Haha - sorry, I had transposed it also, and thought the question was minus instead of multiply - sorry about that confusion!!
Time for more coffee... ;-)
- Bill
Subject: RE: Expected Wait Time in ICM Script
Replied by: Mert Guray on 09-06-2011 06:30:04 AM
No problem
I've got it. Thanks.Subject: RE: Expected Wait Time in ICM Script
Replied by: deniz caglar on 09-08-2011 04:53:29 AM
Hi everyone,
We try this algoritm in one of our skillgroups but it seems the calculated results are not correct.
I wrote this calculated value to a CallVariable and got the seconds i wait in the queue, results are very different.
For example it calculated 52 sc but i my call was handled in 19 sc, and in the other one, it calculated 249 sc, but call was handled in 67sc.
I have two points now.
I think this algoritm doesnt consider the order of the queue. I mean if i am in the queue for 10 minutes and another customer is in the queue for 1 one minute, this algoritm reads the same value to both of us. We dont only read this value to the customer at the first time he is in the queue, but we read it periodically may be once a minute. So may be in the second interval of the second customer that means customer is in the queue for 1 minute, and the tenth interval of the first customer that means he is waiting for 10 minutes, both of them will be informed with the same wait time value. But the first customer will be handled before the second one.
And the other point is; in the Networkers 2011 last month, in the CVP session, they said that there would be a fix in 8.5 for Courtesy Callback. If i am not wrong, in 8.0 if the customer doesnt accept the callback option, then he will be in the last order of the queue, so they dont calculate the order of the customer in the queue dynamically every time. This issue also makes me suspicious about this algoritm.
Do you have any idea or do you use this algoritm in production for giving information to your customers?
Thanks,
Deniz
Subject: RE: Expected Wait Time in ICM Script
Replied by: Paul Antinori on 30-08-2011 04:47:48 PM
The problem we saw in the 8.0 release with EWT is that the callback time is fixed at the time when the caller initially enters the queue.
That implementation does not dynamically update the callback time when the call center call handle rate changes, which renders the ccb feature less effective especially when the callback time is long.
The enhancement in 8.5 release which is available as a patch, is that we can update the EWT (dequeue rate) for ALL of the calls in the CCB queue instead of for just the one call.
The goal of the enhancement was to get the "average callback queue time" after a callback within an agreed service level (default is 1 min).
We do not want to call back callers too early so that they would need to wait in queue for a long time.
We also do not want to call back too late so that agents may be waiting.
So we try get a large percentage of the calls handled within the service level SLA, and the remainder of the calls are falling outside this service level SLA, like a bell curve.
The randomness of call handle time is outside the control of CVP, of course. It could be why you are seeing odd results with your testing.
CVP customers could reduce the randomness of call handle time through business analysis such as separating calls for sales and technical supports into different queues.
Subject: Re: New Message from Paul Antinori in Customer Voice Portal (CVP) - General
Replied by: Janine Graves on 30-08-2011 10:24:54 PM
Hi Paul,
For purposes of recalculating EWT, what is Cisco using as the formula
for EWT while a caller is waiting for callback? And is there any way to
modify this formula?
Thanks, Janine
On 8/30/2011 4:47 PM, Cisco Developer Community Forums wrote:
> Paul Antinori has created a new message in the forum "General
> Discussion - All Versions":
>
> --------------------------------------------------------------
> The problem we saw in the 8.0 release with EWT is that the callback
> time is fixed at the time when the caller initially enters the queue.
> That implementation does not dynamically update the callback time when
> the call center call handle rate changes, which renders the ccb
> feature less effective especially when the callback time is long.
> The enhancement in 8.5 release which is available as a patch, is that
> we can update the EWT (dequeue rate) for ALL of the calls in the CCB
> queue instead of for just the one call.
>
> The goal of the enhancement was to get the "average callback queue
> time" after a callback within an agreed service level (default is 1 min).
> We do not want to call back callers too early so that they would need
> to wait in queue for a long time.
> We also do not want to call back too late so that agents may be waiting.
> So we try get a large percentage of the calls handled within the
> service level SLA, and the remainder of the calls are falling outside
> this service level SLA, like a bell curve.
> The randomness of call handle time is outside the control of CVP, of
> course. It could be why you are seeing odd results with your testing.
> CVP customers could reduce the randomness of call handle time through
> business analysis such as separating calls for sales and technical
> supports into different queues.
> --
> To respond to this post, please click the following link:
>
> <http://developer.cisco.com/web/cvp/forums/-/message_boards/view_message/4431792>
>
> or simply reply to this email.
--
Janine Graves
Subject: RE: Expected Wait Time in ICM Script
Replied by: Xin Yang on 30-08-2011 03:35:13 PM
Hi Deniz,
This is interesting. How about minus current queue length to get the EWT in middle of the queue?
Thanks
James
Subject: RE: Expected Wait Time in ICM Script
Replied by: Paul Antinori on 31-08-2011 08:55:10 AM
Hi Janine,
The default ICM script for CCB has the exact same formula as above. Of course you can customize it for the customer call center.
ValidValue(((XXXX.RouterCallsQNow+1)*(ValidValue(XXXX.AvgHandledCallsTimeTo5,20))/max(XXXX.Ready,(XXXX.TalkingIn+XXXX.TalkingOut+XXXX.TalkingOther))),100)
For 8.5 ES it appears in 2 places in the script to allow for more frequent updating via ged125 messages of the CCB Queue EWT calculation.
Paul
Subject: RE: Expected Wait Time in ICM Script
Replied by: Sethuramalingam Balasubramanian on 03-01-2012 12:47:27 AM
Hi Paul,
Can we know in which version of UCCE/ICM this fix is available or what is the appropriate ES version to this fix?
Thanks!
-Sethu
Subject: RE: Expected Wait Time in ICM Script
Replied by: Brenda McAleenan on 12-09-2013 07:32:00 AM
Hi All
I've made a couple of amendments to this to allow consideration of available agents (without having to use another IF node). If agents are available we'll always set EWT to 0 as they'll be answered straight off.
if(SkillGroup.%1%.Avail>=1,0,ValidValue(((XXXX.RouterCallsQNow+1)*(ValidValue(XXXX.AvgHandledCallsTimeTo5,20))/
max(XXXX.Ready,(XXXX.TalkingIn+XXXX.TalkingOut+XXXX.TalkingOther))),100)
Additionally if you queue to multiple skills you can tweak the formula to suit this. First thing is that you'll only want to check RCQ in one skill (as it should be queuing to all)
if(SkillGroup.%1%_Lvl1.Avail+SkillGroup.PrePay_Lvl2.Avail>=1,0,ValidValue(((SkillGroup.%1%.RouterCallsQNow+1)*(ValidValue((ValidValue(SkillGroup.%1%.HandledCallsTimeTo5,0)+ValidValue(SkillGroup.%2%.HandledCallsTimeTo5,0))/(ValidValue(SkillGroup.%1%.CallsHandledTo5,0)+ValidValue(SkillGroup.%2%.CallsHandledTo5,0)),1))/max((SkillGroup.%1%.Ready+SkillGroup.%2%.Ready),(SkillGroup.%1%.TalkingIn+SkillGroup.%1%.TalkingOut+SkillGroup.%1%.TalkingOther+SkillGroup.%2%.TalkingIn+SkillGroup.%2%.TalkingOut+SkillGroup.%2%.TalkingOther))),100))
You can't use AvgHandledCallsTimeTo5 here anymore as you'd be averaging an average so I've replaced with the manual calculation using HandledCallsTimeTo5/CallsHandledTo 5 looking to both skills. You also need to use the ValidValue logic as these could return NULL for some of the objects (I've noticed CallsHandledTo5 definitely skews this) and therefore default to 20 from the original ValidValue in the formula. I've also found that putting them in a custom function is much easier as they can be reused or updated much easier that way.